Among the principles that drive journalists is their obligation to tell the truth. This principle encompasses two major foundations—accuracy and fact verification. Journalists need to verify all facts to ensure the accuracy of their reports. And it seems that machines are becoming quite adept at journalism as they rely heavily on data and verified facts.
You read that right. Machines can be journalists, too. Please take the time to read the sports news article below, taken from the Washington Post.
The news article above is an example of automated journalism. Automated journalism or “robot journalism” is being employed not only by Washington Post but also by other major news organizations, including Associated Press (AP), USA Today, and Yahoo!
Let’s find out more about automated journalism and how it is changing the newsroom dynamics.
What is Automated Journalism?
Automated journalism refers to the production of news articles by machines or computer programs. It is possible thanks to artificial intelligence (AI) algorithms that allow an application to perform the following functions:
- Scanning enormous amounts of data provided
- Selecting an article template or structure
- Inserting key data points such as places, names, scores, and rankings
And just like that, a news article is then ready for publishing. A computer program, such as Heliograf, which Washington Post uses, can do all these in a matter of seconds.
Reuters also recently took automated journalism to the next level by using deep fake technology to produce a video report without human intervention. Watch the video below, which shows a sports presenter giving the highlights of a game.
Deep fake allows Reuters to have an AI-generated version of the sports presenter read the news script that was written by a machine.
Automated Journalism: Churning Hundreds of News Articles per Day
Heliograf was able to generate 300 news articles when it was launched to cover the Rio Olympics. Aside from news stories, it also created tweets for Washington Post’s Twitter account.
Automated Insights’s Wordsmith, another “writing robot,” creates 1.5 billion articles per year, which is way more than what a team of human writers can accomplish. Wordsmith has written more than 50,000 news stories for AP, ranging from corporate earnings reports to sporting events recaps.
Computer programs can indeed generate an impressive amount of news stories, giving news organizations more content. It also allows them to cover local stories that are otherwise overlooked.
Automated Journalism: No Need for Human Intervention?
According to Automated Insights, Wordsmith has given journalists more time to pursue more complex stories. While this is true, remember that writing robots rely on a preprogrammed writing structure or template. Humans create these templates, so ultimately, automated journalism still requires human inputs.
Then there is also the data that automated journalists use to generate factual news stories. Humans need to verify the data to ensure accuracy before feeding it to the writing robot.
Finally, there is a need for a human journalist to give machine-generated reports a final assessment. Among the questions that journalists ask is, “Does the story have the right context?”
As automated journalism still relies on human inputs and contextualization, the results are subject to biased datasets and algorithms. A machine-generated news report could reflect the creator’s social and political biases, begging the question, “Who would be accountable for that?”
Automated journalism is starting to become a norm, at least in fields that rely heavily on data such as in sports and finance. At present, its impact can be felt when human journalists are given more time to explore other news angles and stories instead of doing repetitive work. News organizations are also able to publish more content and are reaching a wider audience.
Ultimately, though, we are more concerned about automated journalism’s impact on accountability. Human journalists are bound by their commitment to the truth, conscience, and loyalty to the citizens. How can these ethical principles be cascaded to robot journalists? It looks like ethical AI has to be extended to include automated journalism.